Substance use during adolescence is an oft-studied phenomenon, but this research generally fails to take an ecological perspective on etiology. Schools are a primary context of socialization during adolescence, and understanding how school composition matters for substance use is critical for prevention efforts. Promoting school diversity has been a major legislative goal, but the unintended public health consequences of such policies are often ignored-diversity has empirically established academic benefits, yet it is not without its challenges, particularly regarding the socioemotional well-being of children and adolescents whose lack of demographic fit with their schools puts them at risk for social marginalization. Whether this demographic misfit (i.e., having few same-race/ethnic or same-socioeconomic peers in school) is risky for substance use has yet to be explored, although both theory and empirical evidence suggests that it might. The general goal of this project, therefore, is to examine whether, why, and when students who do not have a critical mass of same-race/ethnicity peers or peers of similar SES in school are more likely to drink and use drugs. Here, I use data from Add Health to explore three specific areas of inquiry. First, I will identify adolescents who are at the numeric margins of their schools both racially/ethnically and socioeconomically and compare their substance use to that of adolescents who have greater representation of same-demographic peers. Such research will highlight the potential unintended health risks of major academically-focused school reforms. Second, I will test two mechanisms by which marginalization might influence substance use: a) whether marginalization initiates feelings of misfit that, in turn, contribute to adolescents' substance use and b) whether the link between marginalization and substance use is stronger for students in schools and peer groups in which substance use is more normative. Third, the project will explore the extent to which the marginalization threshold (defined as 15% or more same-demographic peers) effectively captures the critical mass necessary for protection against substance use and lack of fit. Although the National Academy of Education recommends the 15% same- demographic peer threshold to protect against the harmful effects of marginalization, their report acknowledges that this estimate needs empirical validation. As a departure from previous, small-scale studies that explore the critical mass question, this project uses a large, nationally representative sample to empirically identify the critical mass needed to protect against social marginalization. Early substance use and abuse exert pernicious effects across the life course, and this project has the potential to expand our understanding of the implications of school composition for such risky health behaviors. By elucidating the mechanisms by which marginalization affects substance use, the project will highlight critical points of intervention, and by identifying the contextual antecedents of early substance use, the project will inform educational policy efforts that seek to better promote the full academic benefits of diversity in America's public schools.